- Email: [email protected]
Deterioration Index (DI): a suggested criterion for assessing the health of coral communities
Marine Pollution Bulletin 48 (2004) 954–960 www.elsevier.com/locate/marpolbul
Deterioration Index (DI): a suggested criterion for assessing the health of coral communities O. Ben-Tzvi
a,b,*
, Y. Loya b, A. Abelson
a
a
The Institute for Nature Conservation Research, The George S. Wise Faculty of Life Sciences, Tel-Aviv University, 69978 Tel-Aviv, Israel b Department of Zoology, The George S. Wise Faculty of Life Sciences, Tel-Aviv University, 69978 Tel-Aviv, Israel
Abstract The extensive deterioration of coral reefs worldwide highlights the importance of creating efficient monitoring methods to best assess their state of health. At present, several suggested parameters serve such indicators. None of these, however, is well accepted as reliably representing reef community health. In the present study we examine a new approach based on the ratio between mortality and recruitment rates of branching corals, which we term ÔDeterioration Index’ (DI). It aims at providing a quantitative indication of the state of health of reef-building coral communities. The method was developed and tested on 16 coral communities on artificially laid rocks along the coast of Eilat, Red Sea (Gulf of Aqaba). In contrast to frequently used indices (i.e. mortality rate, abundance and species richness), which did not demonstrate a consistent result in comparing disturbed vs. undisturbed coral communities, the DI revealed significant differences between these communities. Our results suggest that the use of the DI may enable the detection of disturbed coral communities in one instance monitoring, where the other parameters had failed. The DI, therefore, may provide a comparable quantitative assessment of the deterioration process and its intensity in a coral community. We propose the DI approach as an efficient and applicable tool for coral reef monitoring. 2003 Elsevier Ltd. All rights reserved. Keywords: Coral reef; Deterioration; Monitoring; Recruitment rate; Mortality rate; Red Sea
1. Introduction Coral reefs are deteriorating globally due to both direct and indirect effects of human activities (e.g. Birkeland, 1996; Wilkinson, 2000). This extensive deterioration underlines the importance and urgency of determining versatile and efficient methods for monitoring the state of reef health (McKenna et al., 2001). At present, however, there is no well-accepted reliable means of indication for a ‘‘healthy reef’’ (Pennisi, 1997; Jameson and Erdmann, 2001; Risk et al., 2001). Several parameters have been suggested as indicators of the state of health of a coral community (Gulko, 2001; *
Corresponding author. Present address: The Interuniversity Institute of Marine Sciences in Eilat, POB 469, Eilat 88103, Israel. Tel.: +972-54-846383; fax: +972-86374329. E-mail address: [email protected] (O. Ben-Tzvi). 0025-326X/$ - see front matter 2003 Elsevier Ltd. All rights reserved. doi:10.1016/j.marpolbul.2003.11.022
Jameson and Erdmann, 2001), particularly: the ratio between living and dead corals (Yap, 1986), or the Mortality Index (Gomez et al., 1994); the size–frequency distribution of corals, which aims to assess the state of a reef by estimating the proportion of small corals, which in turn may indicate recruitment rate to the reef (in healthy coral communities, the frequency of the smallest size groups are expected to be the highest; Bak and Meesters, 1998; Meesters et al., 2001); and the parameter of live coral cover in the reef (Loya, 1972). None of these parameters, however, is accepted as an indication that reliably represents reef community health. Mortality and recruitment rates can vary temporally due to cat-astrophes (Yap, 1986) or normal fluctuations (Hughes et al., 1999), and spatially because of differences in coral survivorship in different reef patches (Miller et al., 2000). Live cover and species richness differ normally between different types of coral reefs (Pennisi, 1997), and can even
O. Ben-Tzvi et al. / Marine Pollution Bulletin 48 (2004) 954–960
differ between patches in the same reef. Moreover, high live cover may misleadingly indicate a healthy reef whereas, in fact, the reef may be undergoing deterioration due to either high mortality and/or low recruitment rates. Therefore, we should be cautious in relying on the above-mentioned parameters as ‘‘health’’ indicators (Birkeland, 1996). Changes over time may provide more reliable indications of the state of the reef (Birkeland, 1996; Pennisi, 1997), but this requires long-term monitoring and thus does not answer the urgent need for immediate large-scale monitoring. A broad-based monitoring approach should meet several requirements: (1) it should enable reliable comparison between different reef types (e.g. reefs of different live cover); (2) it should be simple to apply, including by non-scientific personnel (e.g. recreational divers); (3) it should not require repeated serial surveys, but be able to provide some indication of the state of health of the reef from a single survey event; (4) it should provide an indication of the trend of reef health (developing or deteriorating) rather than only the current state of the reef; and (5) it should provide a quantitative, or at least semiquantitative, indication of the reef state, to enable comparisons between distinct reefs of different characters. The present study suggests a new approach that may help to overcome some of these problems. This approach is based on a suggested parameter, termed ÔDeterioration Index’ (DI), which is the ratio between mortality and recruitment rates of branching corals. Our aim was to test the possible applicability of the DI measure by examining it on coral communities at relatively early successional stages and compare its validity to some widely used indices in coral reef ecology.
955
2. Methods We compared shallow water coral communities of artificially laid rock structures (e.g. breakwaters, road embankments) at 16 different stations in 7 sites along the 12 km coastline of Eilat (Gulf of Aqaba, Fig. 1). At each site one to four stations were chosen based on the structure, rock size and physical diversity of the site. The history of disturbances at each site, from its construction until recent days, is known. This knowledge served in determining the locations and timing of stress events on the corals. Based on this information, we expected to find disturbed communities at the port station (E1) located under the potassium and phosphate quay of Eilat’s port and at the Naval Base station (N1). We also expected to observe traces of changes in the environment, whether the permanent effects of mass production of fish in the farms near the Peace Lagoon stations (P1, P2 and P3), which began four years before our survey, or the temporary effects of construction works in the sea at the Dekel Beach stations (D1, D2 and D3), which took place two years before our survey. Lower stress levels were expected at the southern sites (Lighthouse stations) from which pollution sources are relatively distant and where recreational activities are limited. Each station was censused by three fixed 10 · 0.6 m belt transects. Transects were parallel to shore line in shallow water (2.5–5 m depth). In each transect all living and dead stony corals were counted. Living corals were identified to the species level. The longest colony diameter was measured in living branching and massive corals. The data were used to calculate: (1) coral density, the number of colonies per m2 ; (2) species diversity
Fig. 1. A. Regional map and B. Study sites along Eilat’s coast (Gulf of Aqaba Red Sea): (P) Peace Lagoon, three stations; (Y) Yachts Marina, two stations; (N) Naval Base, one station; (D) Dekel Beach, three stations; (E) Eilat’s port, two stations, one under the phosphate quay and one 100 m southern at the Dolphin Reef; (T) (Tur-Yam) old marina, one station; (L) lighthouse beach, four stations. Major pollution sources are marked with *.
O. Ben-Tzvi et al. / Marine Pollution Bulletin 48 (2004) 954–960
indices: H0 , Shannon and Weaver (1948) and evenness H0 /H0max (Pielou, 1966) and D, Simpson’s index (1949); (3) species richness, the number of species per transect; (4) the proportion of dead branching corals from the total living and dead branching corals; and (5) the proportion of small branching corals, the fraction of corals up to 3 cm (0–3 cm) in diameter from all living branching corals. One way ANOVA (p < 0:05) was used to examine the differences between the stations in all the above-mentioned parameters. Data were tested for homogeneity of variance (Cochran test, STATISTICA release 6) and transformed when required (arcsine for proportions and percentages and log for others). Post hoc tests (Tukey HSD, STATISTICA release 6) were used to indicate stations responsible for the differences found in each parameter. Size–frequency distribution histograms (SFD) of the branching corals at each station were plotted. Corals were divided into three cm groups (0–3, 4–6, etc.). We considered 3 cm as a rough estimate of the average annual growth of a branching colony. For the DI, the ratios between the proportions of dead branching corals and small branching corals at each station were calculated using the formula: DC SC DI ¼ ð1Þ DC þ LC LC where the mortality rate is the proportion of dead colonies (DC) out of the total number of living and dead corals (DC + LC; Gomez et al., 1994); and the recruitment rate is given by the ratio between the number of the smallest detectable living corals (up to 3 cm; SC) and total number of all living corals (LC). In cases where there were no last year recruits (zero colonies in the smallest size group) there is a technical problem of infinite DI. Bigger sample sizes may help to overcome this problem. Otherwise, a mathematical Ôsolution’ is to change the zero value to 0.1. Such a change still results in a very high DI value, but enables the calculation of average DIs and their statistical testing (yet, mathematical corrections have been applied in some transects at several stations). DIs of the different stations were compared using one way ANOVA (p < 0:05) and Tukey post hoc test. Six stations (P2, M1, M2, N1, D2 and L3) were monitored two more times by the DI method. The repeated monitoring was done 6 and 12 months after the original study, in order to test potential seasonal effects on the DI values and to examine if the trends shown by the initial monitoring term were systematic.
3. Results The DI values indicated significant differences between the stations (ANOVA, p < 0:05; Fig. 2). Similar
80 70 60
Deterioration Index
956
50 40 30 20 10 0 P1
P2
P3
Y1
Y2
N1
D1
D2
D3
E1
E2
T1
L1
L2
L3
L4
station
Fig. 2. The Deterioration Indexes (DIs) of the 16 study stations. Peace Lagoon, south breakwater (P1). Peace Lagoon, end of south breakwater (P2). Peace Lagoon, east breakwater (P3). Yachts marina, south side of breakwater (Y1). Yachts marina, east side of breakwater (Y2). Naval port (N1). Dekel beach, plain (D1). Dekel beach, slope (D2). Dekel beach, big rocks (D3). Port (E1). Dolphin Reef (E2). Tour Yam (T1). Lighthouse beach, small size rock (L1). Lighthouse beach, medium size rock (L2). Lighthouse beach, big size rock (L3). Lighthouse beach, southern slope (L4). The DI value of P4 is infinite since no recruits have been found. Bars are ±SD.
results were obtained by the other indices examined, namely: mortality rate of branching corals, recruitment rate of branching and massive corals, coral abundance, species richness, species diversity and evenness (Simpson and Shannon Weaver indices). However, Tukey post hoc tests indicate that different stations are responsible for the significant differences in each parameter (Fig. 3). Moreover, only the DI results pointed out one group of stations (P1, P2 and P3) as the source for the significant p value, while all the other examined parameters did not show a clear picture. The plotted SFD histograms of branching corals show the naturally expected distribution (Bak and Meesters, 1998) at most stations (e.g. Fig. 4C) with natural fluctuations (Hughes et al., 1999), such as the decline seen in the fifth column of Fig. 4C that was found at most stations. At the port (E1), where coral abundance and recruitment are very low, this fifth column is absent (Fig. 4B). At D1 the third column is lower than expected (in a healthy distribution) and in D2 the second column is the low one. Low recruitment was observed in the histograms of P (e.g. P1 Fig. 4A) during the three to four years prior to our sampling and at E1 two years before sampling. Average DI values obtained at the six stations which were re-monitored are shown at Table 1. The ANOVA tests did not reveal significant changes within any station (p < 0:05), although some improvement (lower DI) was observed at Y1 and D2. Likewise, the SFD
O. Ben-Tzvi et al. / Marine Pollution Bulletin 48 (2004) 954–960
957
Fig. 3. The results of Tukey post hoc tests of: recruitment rate; mortality rate; number of species per transect; number of colonies per m2 and DI. * Indicates significant differences between stations (p < 0:05). Each station is compared to all other stations (vertical versus horizontal).
Frequency %
P1 (n = 244)
E1 (n = 35)
L1 (n = 157)
30
30
30
25
25
25
20
20
20
15
15
15
10
10
10
5
5
5
0
3 6 9 12 15 18 21 24 27 30 33 36 39 42
0
3 6 9 12 15 18 21 24 27 30 33 36 39 42
0 3
6 9 12 15 18 21 24 27 30 33 36 39 42
Max colony diameter of size group (cm)
(A)
(B)
(C)
Fig. 4. Size–frequency distribution of branching corals at three stations along the coast of Eilat.
Table 1 Average DI values (±standard deviation) obtained at six selected sites in repeated monitoring along one year SITE
Oct-01
May-02
Oct-02
P2 Y1 Y2 N1 D2 L3
25.52 ± 15.6 1.20 ± 0.69 0.2 ± 0.04 7.81 ± 0.12 2.83 ± 1.5 1.07 ± 0.48
44.43 ± 14.1 0.85 ± 0.41 0.36 ± 0.07 5.91 ± 1.77 1.51 ± 0.31 1.07 ± 0.09
39.80 ± 15.7 0.66 ± 0.1 0.39 ± 0.08 5.41 ± 1.46 1.47 ± 0.35 0.72 ± 0.05
histograms of all stations showed that the tendencies found in the first monitoring were consistent. P2’s histogram shows that the first five size groups are smaller than the sixth which indicates on five years of low recruitment, whereas, at station D2 the high recruitment that started after the construction works is continuing and the relatively low abundance size group that was the second in 2001 is the third in 2002 (Fig. 5).
958
O. Ben-Tzvi et al. / Marine Pollution Bulletin 48 (2004) 954–960 D2 Oct. 2001(n=292)
Frequency %
40
D2 Oct. 2002 (n=272)
40
40
35
35
35
30
30
30
25
25
25
20
20
20
15
15
15
10
10
10
5
5
5
0
0 3
9
15
21
27
33
39
45
51
3
9 15
21
27
33
39
45
51
0
64.4% D1 Oct 2001 (239 colonies)
3
9 15
21
27
33
39
45
51
Max colony diameter of size Group (cm)
Fig. 5. Size–frequency distribution of branching corals at station D2 in 2001 (left) and 2002 (middle) and D1(right). The relatively low frequency of corals in the 2nd size group at D2 in 2001 shows the low recruitment caused by works at this station 2 years before. A year later, in 2002 the low frequency is seen at the same station in the 3rd size group. At D1 the 3rd size group shows relatively low frequency due to the works that occurred there 3 years before.
4. Discussion The idea underlying the DI is that certain reefs may experience different mortality and recruitment rates from other reefs, and yet can show a stable state. Therefore, considering mortality and recruitment independently may lead to an erroneous picture of the reef’s true state. The DI, however, takes into account both parameters, thus, when high mortality rates and low recruitment rates coincide at the same reef, the community is likely to undergo deterioration. By contrast, if mortality rate is low and recruitment rate is high the coral community is in a healthy state and on occasions may even undergo rehabilitation following a major disturbance (e.g. bleaching, hurricane etc.). In contrast to the commonly used parameters (i.e. mortality rate, abundance, species richness and species diversity), the DI revealed significant differences (p < 0:05; Figs. 2 and 3) between all the stations of one site, P (P1, P2 and P3), and the rest of the stations on the artificially laid rocks. The DIs at the P stations were found to be significantly higher, indicating a deteriorating community. The same high level of DI found at P2 (Table 1) in the repeated monitoring shows that the deterioration of the coral community is continuing. Use of the DI, therefore, enabled us to identify disturbed sites, as opposed to the other widely used parameters, which have failed to reveal any clear significant differences. The Peace Lagoon is known to be exposed to several disturbances, the major one of which is a high nutrient level, derived from three potential sources (Abelson et al., 1999; Shlesinger and Lazar, 2001): (1) fish-rearing farms in which more than 4,000 tons of fish food have been released into the sea annually (Atkinson et al., 2001) over the last four years. These farms are located in the sea 500 m from the Peace Lagoon; (2) pollution of fertilizers from the artificial lagoon; and (3) accidental spills of Eilat’s sewage into the sea following failures of the sewage system. In addition to the high nutrient levels, the P site is also exposed to high
sedimentation from a terrestrial source. Sediment is often brought by the frequent northern wind and causes high turbidity. In addition to the P stations, there were several other stations that revealed high DI values (N1, E1 and D2 Fig. 2). These stations are among those which were considered to be stressed due to pollution and other anthropogenic disturbances. These three stations were not found to be significantly different from other stations in the Tukey test. This is due to the high standard deviations that were obtained at the Peace Lagoon stations and at N1 and E1 (Fig. 2). Very low recruitment at the Peace Lagoon, low coral abundance at E1 and low branching coral abundance at N1 yielded large differences in the proportions of small corals which in turn caused the changes in the DI values. In spite of this the average DI values expressed the expected stress at all these stations. DI was also found sensitive enough to show clearly the differences between Y1 (DI ¼ 1.2) and Y2 (DI ¼ 0.2). Our observations noted a problem at Y1, where the highest diversity, species richness and live cover are found. It was found that the sea urchin (mainly Diadema sp.) density is low and turf algae are dense. We do not know what differentiates this station from station Y2, only 20 m away, where sea urchin abundance was found to be eight times higher in October 2001. However, the abundance of Diadema sea-urchins at Y1 was doubled in 2002. The significantly lower DI values at this station in this year are assumed to be the result of higher recruitment which can be the direct effect of the higher sea urchins density (Sammarco, 1980). When the coral-community state is stable, the DI value is expected to be low regardless of live cover and species diversity. However, when the community is in decline, the DI will be high whether coral abundance and species richness are high or low, and whatever the species assemblage may be. The DI manifestation of the community state can be seen from the size–frequency distribution of corals in our study sites. The size–frequency
O. Ben-Tzvi et al. / Marine Pollution Bulletin 48 (2004) 954–960
distribution of branching corals at P1 (Fig. 4A) shows a low proportion of small colonies (up to 9 cm in diameter), which may indicate very low recruitment during the last 3–4 years. A similar distribution pattern was obtained from station E1 (Fig. 4B). However, mortality rate at station E1 was found to be low compared with P1. While coral abundance at E1, therefore, is very low, as a consequence of extremely low recruitment rate, the coral-community state is stable, since mortality rate is low as well. These differences between sites E1 and P1 are reflected in their DI values (Fig. 2). However, the DI cannot reveal the history of the examined coral community. Some indication of this can be determined from the SFD of corals, which can trace changes in mortality and recruitment of the examined community. For instance, the histograms of most of our studied communities reveal a decline in the size range of 12–15 cm (Fig. 4 B and C). This decline may indicate low recruitment along the Eilat coast around four to five years ago. Likewise, it can be clearly seen that at P1, recruitment was very low over the last three to four years (Fig. 4A), which can explain the very high DI value received from the P1 station. In this particular case the data also point to a possible link between the sharp decline in recruitment (and the consequent reef deterioration) and the dramatic increase in food supply to the fish farms, a major pollution source located 500 m from the Peace Lagoon (according to the Israeli fisheries department, the fish production increased from 450 tons in 1994 to 2400 tons in 1999 with the major increase of 87% from 1996 to 1997). The very low recruitment at this site continued in 2002 as shown by the SFD of P2 from October 2002, as was predicted by the DI of 2001, underlined by the DI of 2002 and expected due to the continuity in the flow of pollution to the sea. The port station (E1) has been exposed frequently to phosphate pollution due to leakage from the phosphate quay (under which it is located) since its construction about 35 years ago. This dictates the development of a meager coral community, which is not changing. However, since the community state is relatively stable the DI is lower than that of the Peace Lagoon stations (Fig. 2). It should be noted that the three cm range of the group sizes was chosen without having sufficient data on the growth rates of all branching corals. Well-dated events like the construction works at site D corroborate this figure. The construction works took place close to station D1 about three years prior to our first monitoring, and are reflected by a clear decline in the frequency of the third group size (Fig. 5). The construction work continued nearby station D2, where they were finished about 2 years prior to our monitoring. Here the construction effect on the coral community is reflected by a decline in the second group size (Fig. 5). One year later in the SFD of the same station from October 2002 the decline has shifted and it is seen in the third column.
959
At this stage we are focusing the DI calculations on the recruitment and mortality rates of branching corals. This approach of using functional groups follows Knowlton’s (2001) approach claiming that functional groups can be used for reef monitoring purposes. The functional group of Ôbranching corals’ was selected for the DI for several practical reasons. As long as the branching structure of dead branching corals has not been eroded, they can be easily detected and counted as ‘‘recently dead branching corals’’. In contrast, dead massive corals are difficult to detect shortly after death. Moreover, if detected, it is impossible to estimate the time which elapsed since mortality, as the general shape of massive colonies can be maintained for a long period. Branching corals usually recruit in larger numbers and are more sensitive to disturbances. These type-2 corals (fast growing with medium resistance to disturbances) are better indicators for the whole coral-community state than corals that are more sustainable, like most of the massive corals, which are type-1 (Hughes, 1985). Furthermore, in order to take into account both massive and encrusting corals in the DI calculations, a preparatory study is required for determination of erosion and accretion rates so that dead corals can be recognized and their mortality time estimated. The most significant contribution of the DI may lie in its ability to indicate the health trend of a given coralcommunity (i.e., whether it is developing or deteriorating), rather than reflecting the current community state alone, as indicated by other parameters like live cover, community structure and species diversity. Finally, the DI may provide a comparable quantitative indication of the existing community deterioration process and its intensity. All these can be obtained from data collected by volunteer lay divers, since no coral identification is required. Based on these advantages, we suggest the DI as a reliable criterion, and efficient tool for assessing the health of coral communities. Acknowledgements This study was part of the Red Sea Marine Peace Park (RSMPP) program funded by USAID Program and the Raynor Chair for Environmental Conservation Research. The authors thank the Director and staff of the Marine Biology Laboratory at Eilat for their hospitality and the use of lab facilities. We also thank Naomi Paz and Karen Madmoni for their editorial assistance. References Abelson, A., Shteinman, B., Fine, M., Kaganovsky, S., 1999. Mass transport from pollution sources to remote coral reefs in Eilat (Gulf of Aqaba, Red Sea). Mar. Pollut. Bull. 38, 25–29. Atkinson, M.J., Birk, Y., Rosenthal, H., 2001. Evaluation of pollution in the Gulf of Eilat. Report for the Ministries of Infrastructure,
960
O. Ben-Tzvi et al. / Marine Pollution Bulletin 48 (2004) 954–960
Environment and Agriculture. Published by the Chief Scientist Office in the Ministry of Agriculture & Rural development, State of Israel. Bak, R.P.M., Meesters, E.H., 1998. Coral population structure: the hidden information of colony size–frequency distribution. Mar. Ecol. Prog. Ser. 162, 301–306. Birkeland, C., 1996. Introduction to Life and Death of Coral Reefs. Chapman & Hall, New York. Gomez, E.D., Ali~ no, P.M., Yap, H.T., Licuanan, W.Y., 1994. A review of the status of Philippine reefs. Mar. Pollut. Bull. 29, 62– 68. Gulko, D., 2001. NCRI Coral Reef Monitoring Methodology Comparison Chart. In: Proc. of the Int. Conf. on Scientific Aspects of Coral reef Assessment, Monitoring and restorationBull. Mar. Sci. 69 (2). Hughes, T.P., 1985. Life histories and population dynamics of early successional corals. In: Proc. Fifth Coral Reefs Symp., vol. 4, pp. 101–106. Hughes, T.P., Baird, A.B., Dinsdale, E.A., Moltschaniwskyj, N.A., Pratchett, M.S., Tanner, J.E., Willis, B.L., 1999. Patterns of recruitment and abundance of corals along the Great Barrier Reef. Nature 397, 59–63. Jameson, J.C., Erdmann, M.V., 2001. Charting a course toward diagnostic monitoring: a continuing review of coral reef attributes and a research strategy for creating coral reef indexes of biotic integrity. Bull. Mar. Sci. 69 (2), 701–744. Knowlton, N., 2001. Who are the players on coral reefs and does it matter. The importance of coral taxonomy for coral reef management. Bull. Mar. Sci. 69 (2), 305–308. Loya, Y., 1972. Community structure and species diversity of hermatypic corals at Eilat, Red Sea. Mar. Biol. 13 (2), 100– 123.
McKenna, S.A., Richmond, R.H., Roos, G., 2001. Assessing the effects of sewage on coral reefs: developing techniques to detect stress before coral mortality. Bull. Mar. Sci. 69 (2), 517–524. Meesters, E.H., Hilterman, M., Kardinall, E., Keetman, M., de Vries, M., Bak, R.P.M., 2001. Colony size–frequency distributions of scleractinian coral populations: spatial and interspecific variation. Mar. Ecol. Prog. Ser. 209, 43–54. Miller, M.W., Weil, E., Szmant, A.M., 2000. Coral recruitment and juvenile mortality as structuring factors for reef benthic communities in Biscayne National Park. Coral Reefs 19, 115–123. Pennisi, E., 1997. Brighter Prospect for the World’s Coral Reefs? Science 277, 491–493. Pielou, E.R., 1966. Species diversity and pattern-diversity in the study of ecological succession. J. Theor. Biol. 13, 370–383. Risk, M.J., Heikoop, J.M., Edinger, R.V., Erdmann, M.V., 2001. The assessment ‘‘toolbox’’: community-based reef evaluation methods coupled with geochemical techniques to identify sources of stress. Bull. Mar. Sci. 69 (2), 443–458. Sammarco, P.W., 1980. Diadema and its relationship to coral spat mortality: grazing, competition and biological disturbance. J. Exp. Mar. Bio. Ecol. 45, 245–272. Shannon, C.E., Weaver, W., 1948. The Mathematical Theory of Communications. University of Illinois Press, Urbana. p. 117. Shlesinger, Y., Lazar, B., 2001. Artificial swimming lagoon in Eilat: monitoring and management. In: Book of Abstracts of the 33rd Anniversary Conference of the IUI Eilat, pp. 131–132. Simpson, E.H., 1949. Measurements of diversity. Nuatur London 163, 668. Wilkinson, C.R., 2000. Status of Coral Reefs of the World 2000. AIMS, Queensland. Yap, H.T., 1986. Bioindication in coral reef ecosystems. Acta Biol. Hung. 37 (1), 55–58.
Deterioration Index (DI): a suggested criterion for assessing the health of coral communities O. Ben-Tzvi
a,b,*
, Y. Loya b, A. Abelson
a
a
The Institute for Nature Conservation Research, The George S. Wise Faculty of Life Sciences, Tel-Aviv University, 69978 Tel-Aviv, Israel b Department of Zoology, The George S. Wise Faculty of Life Sciences, Tel-Aviv University, 69978 Tel-Aviv, Israel
Abstract The extensive deterioration of coral reefs worldwide highlights the importance of creating efficient monitoring methods to best assess their state of health. At present, several suggested parameters serve such indicators. None of these, however, is well accepted as reliably representing reef community health. In the present study we examine a new approach based on the ratio between mortality and recruitment rates of branching corals, which we term ÔDeterioration Index’ (DI). It aims at providing a quantitative indication of the state of health of reef-building coral communities. The method was developed and tested on 16 coral communities on artificially laid rocks along the coast of Eilat, Red Sea (Gulf of Aqaba). In contrast to frequently used indices (i.e. mortality rate, abundance and species richness), which did not demonstrate a consistent result in comparing disturbed vs. undisturbed coral communities, the DI revealed significant differences between these communities. Our results suggest that the use of the DI may enable the detection of disturbed coral communities in one instance monitoring, where the other parameters had failed. The DI, therefore, may provide a comparable quantitative assessment of the deterioration process and its intensity in a coral community. We propose the DI approach as an efficient and applicable tool for coral reef monitoring. 2003 Elsevier Ltd. All rights reserved. Keywords: Coral reef; Deterioration; Monitoring; Recruitment rate; Mortality rate; Red Sea
1. Introduction Coral reefs are deteriorating globally due to both direct and indirect effects of human activities (e.g. Birkeland, 1996; Wilkinson, 2000). This extensive deterioration underlines the importance and urgency of determining versatile and efficient methods for monitoring the state of reef health (McKenna et al., 2001). At present, however, there is no well-accepted reliable means of indication for a ‘‘healthy reef’’ (Pennisi, 1997; Jameson and Erdmann, 2001; Risk et al., 2001). Several parameters have been suggested as indicators of the state of health of a coral community (Gulko, 2001; *
Corresponding author. Present address: The Interuniversity Institute of Marine Sciences in Eilat, POB 469, Eilat 88103, Israel. Tel.: +972-54-846383; fax: +972-86374329. E-mail address: [email protected] (O. Ben-Tzvi). 0025-326X/$ - see front matter 2003 Elsevier Ltd. All rights reserved. doi:10.1016/j.marpolbul.2003.11.022
Jameson and Erdmann, 2001), particularly: the ratio between living and dead corals (Yap, 1986), or the Mortality Index (Gomez et al., 1994); the size–frequency distribution of corals, which aims to assess the state of a reef by estimating the proportion of small corals, which in turn may indicate recruitment rate to the reef (in healthy coral communities, the frequency of the smallest size groups are expected to be the highest; Bak and Meesters, 1998; Meesters et al., 2001); and the parameter of live coral cover in the reef (Loya, 1972). None of these parameters, however, is accepted as an indication that reliably represents reef community health. Mortality and recruitment rates can vary temporally due to cat-astrophes (Yap, 1986) or normal fluctuations (Hughes et al., 1999), and spatially because of differences in coral survivorship in different reef patches (Miller et al., 2000). Live cover and species richness differ normally between different types of coral reefs (Pennisi, 1997), and can even
O. Ben-Tzvi et al. / Marine Pollution Bulletin 48 (2004) 954–960
differ between patches in the same reef. Moreover, high live cover may misleadingly indicate a healthy reef whereas, in fact, the reef may be undergoing deterioration due to either high mortality and/or low recruitment rates. Therefore, we should be cautious in relying on the above-mentioned parameters as ‘‘health’’ indicators (Birkeland, 1996). Changes over time may provide more reliable indications of the state of the reef (Birkeland, 1996; Pennisi, 1997), but this requires long-term monitoring and thus does not answer the urgent need for immediate large-scale monitoring. A broad-based monitoring approach should meet several requirements: (1) it should enable reliable comparison between different reef types (e.g. reefs of different live cover); (2) it should be simple to apply, including by non-scientific personnel (e.g. recreational divers); (3) it should not require repeated serial surveys, but be able to provide some indication of the state of health of the reef from a single survey event; (4) it should provide an indication of the trend of reef health (developing or deteriorating) rather than only the current state of the reef; and (5) it should provide a quantitative, or at least semiquantitative, indication of the reef state, to enable comparisons between distinct reefs of different characters. The present study suggests a new approach that may help to overcome some of these problems. This approach is based on a suggested parameter, termed ÔDeterioration Index’ (DI), which is the ratio between mortality and recruitment rates of branching corals. Our aim was to test the possible applicability of the DI measure by examining it on coral communities at relatively early successional stages and compare its validity to some widely used indices in coral reef ecology.
955
2. Methods We compared shallow water coral communities of artificially laid rock structures (e.g. breakwaters, road embankments) at 16 different stations in 7 sites along the 12 km coastline of Eilat (Gulf of Aqaba, Fig. 1). At each site one to four stations were chosen based on the structure, rock size and physical diversity of the site. The history of disturbances at each site, from its construction until recent days, is known. This knowledge served in determining the locations and timing of stress events on the corals. Based on this information, we expected to find disturbed communities at the port station (E1) located under the potassium and phosphate quay of Eilat’s port and at the Naval Base station (N1). We also expected to observe traces of changes in the environment, whether the permanent effects of mass production of fish in the farms near the Peace Lagoon stations (P1, P2 and P3), which began four years before our survey, or the temporary effects of construction works in the sea at the Dekel Beach stations (D1, D2 and D3), which took place two years before our survey. Lower stress levels were expected at the southern sites (Lighthouse stations) from which pollution sources are relatively distant and where recreational activities are limited. Each station was censused by three fixed 10 · 0.6 m belt transects. Transects were parallel to shore line in shallow water (2.5–5 m depth). In each transect all living and dead stony corals were counted. Living corals were identified to the species level. The longest colony diameter was measured in living branching and massive corals. The data were used to calculate: (1) coral density, the number of colonies per m2 ; (2) species diversity
Fig. 1. A. Regional map and B. Study sites along Eilat’s coast (Gulf of Aqaba Red Sea): (P) Peace Lagoon, three stations; (Y) Yachts Marina, two stations; (N) Naval Base, one station; (D) Dekel Beach, three stations; (E) Eilat’s port, two stations, one under the phosphate quay and one 100 m southern at the Dolphin Reef; (T) (Tur-Yam) old marina, one station; (L) lighthouse beach, four stations. Major pollution sources are marked with *.
O. Ben-Tzvi et al. / Marine Pollution Bulletin 48 (2004) 954–960
indices: H0 , Shannon and Weaver (1948) and evenness H0 /H0max (Pielou, 1966) and D, Simpson’s index (1949); (3) species richness, the number of species per transect; (4) the proportion of dead branching corals from the total living and dead branching corals; and (5) the proportion of small branching corals, the fraction of corals up to 3 cm (0–3 cm) in diameter from all living branching corals. One way ANOVA (p < 0:05) was used to examine the differences between the stations in all the above-mentioned parameters. Data were tested for homogeneity of variance (Cochran test, STATISTICA release 6) and transformed when required (arcsine for proportions and percentages and log for others). Post hoc tests (Tukey HSD, STATISTICA release 6) were used to indicate stations responsible for the differences found in each parameter. Size–frequency distribution histograms (SFD) of the branching corals at each station were plotted. Corals were divided into three cm groups (0–3, 4–6, etc.). We considered 3 cm as a rough estimate of the average annual growth of a branching colony. For the DI, the ratios between the proportions of dead branching corals and small branching corals at each station were calculated using the formula: DC SC DI ¼ ð1Þ DC þ LC LC where the mortality rate is the proportion of dead colonies (DC) out of the total number of living and dead corals (DC + LC; Gomez et al., 1994); and the recruitment rate is given by the ratio between the number of the smallest detectable living corals (up to 3 cm; SC) and total number of all living corals (LC). In cases where there were no last year recruits (zero colonies in the smallest size group) there is a technical problem of infinite DI. Bigger sample sizes may help to overcome this problem. Otherwise, a mathematical Ôsolution’ is to change the zero value to 0.1. Such a change still results in a very high DI value, but enables the calculation of average DIs and their statistical testing (yet, mathematical corrections have been applied in some transects at several stations). DIs of the different stations were compared using one way ANOVA (p < 0:05) and Tukey post hoc test. Six stations (P2, M1, M2, N1, D2 and L3) were monitored two more times by the DI method. The repeated monitoring was done 6 and 12 months after the original study, in order to test potential seasonal effects on the DI values and to examine if the trends shown by the initial monitoring term were systematic.
3. Results The DI values indicated significant differences between the stations (ANOVA, p < 0:05; Fig. 2). Similar
80 70 60
Deterioration Index
956
50 40 30 20 10 0 P1
P2
P3
Y1
Y2
N1
D1
D2
D3
E1
E2
T1
L1
L2
L3
L4
station
Fig. 2. The Deterioration Indexes (DIs) of the 16 study stations. Peace Lagoon, south breakwater (P1). Peace Lagoon, end of south breakwater (P2). Peace Lagoon, east breakwater (P3). Yachts marina, south side of breakwater (Y1). Yachts marina, east side of breakwater (Y2). Naval port (N1). Dekel beach, plain (D1). Dekel beach, slope (D2). Dekel beach, big rocks (D3). Port (E1). Dolphin Reef (E2). Tour Yam (T1). Lighthouse beach, small size rock (L1). Lighthouse beach, medium size rock (L2). Lighthouse beach, big size rock (L3). Lighthouse beach, southern slope (L4). The DI value of P4 is infinite since no recruits have been found. Bars are ±SD.
results were obtained by the other indices examined, namely: mortality rate of branching corals, recruitment rate of branching and massive corals, coral abundance, species richness, species diversity and evenness (Simpson and Shannon Weaver indices). However, Tukey post hoc tests indicate that different stations are responsible for the significant differences in each parameter (Fig. 3). Moreover, only the DI results pointed out one group of stations (P1, P2 and P3) as the source for the significant p value, while all the other examined parameters did not show a clear picture. The plotted SFD histograms of branching corals show the naturally expected distribution (Bak and Meesters, 1998) at most stations (e.g. Fig. 4C) with natural fluctuations (Hughes et al., 1999), such as the decline seen in the fifth column of Fig. 4C that was found at most stations. At the port (E1), where coral abundance and recruitment are very low, this fifth column is absent (Fig. 4B). At D1 the third column is lower than expected (in a healthy distribution) and in D2 the second column is the low one. Low recruitment was observed in the histograms of P (e.g. P1 Fig. 4A) during the three to four years prior to our sampling and at E1 two years before sampling. Average DI values obtained at the six stations which were re-monitored are shown at Table 1. The ANOVA tests did not reveal significant changes within any station (p < 0:05), although some improvement (lower DI) was observed at Y1 and D2. Likewise, the SFD
O. Ben-Tzvi et al. / Marine Pollution Bulletin 48 (2004) 954–960
957
Fig. 3. The results of Tukey post hoc tests of: recruitment rate; mortality rate; number of species per transect; number of colonies per m2 and DI. * Indicates significant differences between stations (p < 0:05). Each station is compared to all other stations (vertical versus horizontal).
Frequency %
P1 (n = 244)
E1 (n = 35)
L1 (n = 157)
30
30
30
25
25
25
20
20
20
15
15
15
10
10
10
5
5
5
0
3 6 9 12 15 18 21 24 27 30 33 36 39 42
0
3 6 9 12 15 18 21 24 27 30 33 36 39 42
0 3
6 9 12 15 18 21 24 27 30 33 36 39 42
Max colony diameter of size group (cm)
(A)
(B)
(C)
Fig. 4. Size–frequency distribution of branching corals at three stations along the coast of Eilat.
Table 1 Average DI values (±standard deviation) obtained at six selected sites in repeated monitoring along one year SITE
Oct-01
May-02
Oct-02
P2 Y1 Y2 N1 D2 L3
25.52 ± 15.6 1.20 ± 0.69 0.2 ± 0.04 7.81 ± 0.12 2.83 ± 1.5 1.07 ± 0.48
44.43 ± 14.1 0.85 ± 0.41 0.36 ± 0.07 5.91 ± 1.77 1.51 ± 0.31 1.07 ± 0.09
39.80 ± 15.7 0.66 ± 0.1 0.39 ± 0.08 5.41 ± 1.46 1.47 ± 0.35 0.72 ± 0.05
histograms of all stations showed that the tendencies found in the first monitoring were consistent. P2’s histogram shows that the first five size groups are smaller than the sixth which indicates on five years of low recruitment, whereas, at station D2 the high recruitment that started after the construction works is continuing and the relatively low abundance size group that was the second in 2001 is the third in 2002 (Fig. 5).
958
O. Ben-Tzvi et al. / Marine Pollution Bulletin 48 (2004) 954–960 D2 Oct. 2001(n=292)
Frequency %
40
D2 Oct. 2002 (n=272)
40
40
35
35
35
30
30
30
25
25
25
20
20
20
15
15
15
10
10
10
5
5
5
0
0 3
9
15
21
27
33
39
45
51
3
9 15
21
27
33
39
45
51
0
64.4% D1 Oct 2001 (239 colonies)
3
9 15
21
27
33
39
45
51
Max colony diameter of size Group (cm)
Fig. 5. Size–frequency distribution of branching corals at station D2 in 2001 (left) and 2002 (middle) and D1(right). The relatively low frequency of corals in the 2nd size group at D2 in 2001 shows the low recruitment caused by works at this station 2 years before. A year later, in 2002 the low frequency is seen at the same station in the 3rd size group. At D1 the 3rd size group shows relatively low frequency due to the works that occurred there 3 years before.
4. Discussion The idea underlying the DI is that certain reefs may experience different mortality and recruitment rates from other reefs, and yet can show a stable state. Therefore, considering mortality and recruitment independently may lead to an erroneous picture of the reef’s true state. The DI, however, takes into account both parameters, thus, when high mortality rates and low recruitment rates coincide at the same reef, the community is likely to undergo deterioration. By contrast, if mortality rate is low and recruitment rate is high the coral community is in a healthy state and on occasions may even undergo rehabilitation following a major disturbance (e.g. bleaching, hurricane etc.). In contrast to the commonly used parameters (i.e. mortality rate, abundance, species richness and species diversity), the DI revealed significant differences (p < 0:05; Figs. 2 and 3) between all the stations of one site, P (P1, P2 and P3), and the rest of the stations on the artificially laid rocks. The DIs at the P stations were found to be significantly higher, indicating a deteriorating community. The same high level of DI found at P2 (Table 1) in the repeated monitoring shows that the deterioration of the coral community is continuing. Use of the DI, therefore, enabled us to identify disturbed sites, as opposed to the other widely used parameters, which have failed to reveal any clear significant differences. The Peace Lagoon is known to be exposed to several disturbances, the major one of which is a high nutrient level, derived from three potential sources (Abelson et al., 1999; Shlesinger and Lazar, 2001): (1) fish-rearing farms in which more than 4,000 tons of fish food have been released into the sea annually (Atkinson et al., 2001) over the last four years. These farms are located in the sea 500 m from the Peace Lagoon; (2) pollution of fertilizers from the artificial lagoon; and (3) accidental spills of Eilat’s sewage into the sea following failures of the sewage system. In addition to the high nutrient levels, the P site is also exposed to high
sedimentation from a terrestrial source. Sediment is often brought by the frequent northern wind and causes high turbidity. In addition to the P stations, there were several other stations that revealed high DI values (N1, E1 and D2 Fig. 2). These stations are among those which were considered to be stressed due to pollution and other anthropogenic disturbances. These three stations were not found to be significantly different from other stations in the Tukey test. This is due to the high standard deviations that were obtained at the Peace Lagoon stations and at N1 and E1 (Fig. 2). Very low recruitment at the Peace Lagoon, low coral abundance at E1 and low branching coral abundance at N1 yielded large differences in the proportions of small corals which in turn caused the changes in the DI values. In spite of this the average DI values expressed the expected stress at all these stations. DI was also found sensitive enough to show clearly the differences between Y1 (DI ¼ 1.2) and Y2 (DI ¼ 0.2). Our observations noted a problem at Y1, where the highest diversity, species richness and live cover are found. It was found that the sea urchin (mainly Diadema sp.) density is low and turf algae are dense. We do not know what differentiates this station from station Y2, only 20 m away, where sea urchin abundance was found to be eight times higher in October 2001. However, the abundance of Diadema sea-urchins at Y1 was doubled in 2002. The significantly lower DI values at this station in this year are assumed to be the result of higher recruitment which can be the direct effect of the higher sea urchins density (Sammarco, 1980). When the coral-community state is stable, the DI value is expected to be low regardless of live cover and species diversity. However, when the community is in decline, the DI will be high whether coral abundance and species richness are high or low, and whatever the species assemblage may be. The DI manifestation of the community state can be seen from the size–frequency distribution of corals in our study sites. The size–frequency
O. Ben-Tzvi et al. / Marine Pollution Bulletin 48 (2004) 954–960
distribution of branching corals at P1 (Fig. 4A) shows a low proportion of small colonies (up to 9 cm in diameter), which may indicate very low recruitment during the last 3–4 years. A similar distribution pattern was obtained from station E1 (Fig. 4B). However, mortality rate at station E1 was found to be low compared with P1. While coral abundance at E1, therefore, is very low, as a consequence of extremely low recruitment rate, the coral-community state is stable, since mortality rate is low as well. These differences between sites E1 and P1 are reflected in their DI values (Fig. 2). However, the DI cannot reveal the history of the examined coral community. Some indication of this can be determined from the SFD of corals, which can trace changes in mortality and recruitment of the examined community. For instance, the histograms of most of our studied communities reveal a decline in the size range of 12–15 cm (Fig. 4 B and C). This decline may indicate low recruitment along the Eilat coast around four to five years ago. Likewise, it can be clearly seen that at P1, recruitment was very low over the last three to four years (Fig. 4A), which can explain the very high DI value received from the P1 station. In this particular case the data also point to a possible link between the sharp decline in recruitment (and the consequent reef deterioration) and the dramatic increase in food supply to the fish farms, a major pollution source located 500 m from the Peace Lagoon (according to the Israeli fisheries department, the fish production increased from 450 tons in 1994 to 2400 tons in 1999 with the major increase of 87% from 1996 to 1997). The very low recruitment at this site continued in 2002 as shown by the SFD of P2 from October 2002, as was predicted by the DI of 2001, underlined by the DI of 2002 and expected due to the continuity in the flow of pollution to the sea. The port station (E1) has been exposed frequently to phosphate pollution due to leakage from the phosphate quay (under which it is located) since its construction about 35 years ago. This dictates the development of a meager coral community, which is not changing. However, since the community state is relatively stable the DI is lower than that of the Peace Lagoon stations (Fig. 2). It should be noted that the three cm range of the group sizes was chosen without having sufficient data on the growth rates of all branching corals. Well-dated events like the construction works at site D corroborate this figure. The construction works took place close to station D1 about three years prior to our first monitoring, and are reflected by a clear decline in the frequency of the third group size (Fig. 5). The construction work continued nearby station D2, where they were finished about 2 years prior to our monitoring. Here the construction effect on the coral community is reflected by a decline in the second group size (Fig. 5). One year later in the SFD of the same station from October 2002 the decline has shifted and it is seen in the third column.
959
At this stage we are focusing the DI calculations on the recruitment and mortality rates of branching corals. This approach of using functional groups follows Knowlton’s (2001) approach claiming that functional groups can be used for reef monitoring purposes. The functional group of Ôbranching corals’ was selected for the DI for several practical reasons. As long as the branching structure of dead branching corals has not been eroded, they can be easily detected and counted as ‘‘recently dead branching corals’’. In contrast, dead massive corals are difficult to detect shortly after death. Moreover, if detected, it is impossible to estimate the time which elapsed since mortality, as the general shape of massive colonies can be maintained for a long period. Branching corals usually recruit in larger numbers and are more sensitive to disturbances. These type-2 corals (fast growing with medium resistance to disturbances) are better indicators for the whole coral-community state than corals that are more sustainable, like most of the massive corals, which are type-1 (Hughes, 1985). Furthermore, in order to take into account both massive and encrusting corals in the DI calculations, a preparatory study is required for determination of erosion and accretion rates so that dead corals can be recognized and their mortality time estimated. The most significant contribution of the DI may lie in its ability to indicate the health trend of a given coralcommunity (i.e., whether it is developing or deteriorating), rather than reflecting the current community state alone, as indicated by other parameters like live cover, community structure and species diversity. Finally, the DI may provide a comparable quantitative indication of the existing community deterioration process and its intensity. All these can be obtained from data collected by volunteer lay divers, since no coral identification is required. Based on these advantages, we suggest the DI as a reliable criterion, and efficient tool for assessing the health of coral communities. Acknowledgements This study was part of the Red Sea Marine Peace Park (RSMPP) program funded by USAID Program and the Raynor Chair for Environmental Conservation Research. The authors thank the Director and staff of the Marine Biology Laboratory at Eilat for their hospitality and the use of lab facilities. We also thank Naomi Paz and Karen Madmoni for their editorial assistance. References Abelson, A., Shteinman, B., Fine, M., Kaganovsky, S., 1999. Mass transport from pollution sources to remote coral reefs in Eilat (Gulf of Aqaba, Red Sea). Mar. Pollut. Bull. 38, 25–29. Atkinson, M.J., Birk, Y., Rosenthal, H., 2001. Evaluation of pollution in the Gulf of Eilat. Report for the Ministries of Infrastructure,
960
O. Ben-Tzvi et al. / Marine Pollution Bulletin 48 (2004) 954–960
Environment and Agriculture. Published by the Chief Scientist Office in the Ministry of Agriculture & Rural development, State of Israel. Bak, R.P.M., Meesters, E.H., 1998. Coral population structure: the hidden information of colony size–frequency distribution. Mar. Ecol. Prog. Ser. 162, 301–306. Birkeland, C., 1996. Introduction to Life and Death of Coral Reefs. Chapman & Hall, New York. Gomez, E.D., Ali~ no, P.M., Yap, H.T., Licuanan, W.Y., 1994. A review of the status of Philippine reefs. Mar. Pollut. Bull. 29, 62– 68. Gulko, D., 2001. NCRI Coral Reef Monitoring Methodology Comparison Chart. In: Proc. of the Int. Conf. on Scientific Aspects of Coral reef Assessment, Monitoring and restorationBull. Mar. Sci. 69 (2). Hughes, T.P., 1985. Life histories and population dynamics of early successional corals. In: Proc. Fifth Coral Reefs Symp., vol. 4, pp. 101–106. Hughes, T.P., Baird, A.B., Dinsdale, E.A., Moltschaniwskyj, N.A., Pratchett, M.S., Tanner, J.E., Willis, B.L., 1999. Patterns of recruitment and abundance of corals along the Great Barrier Reef. Nature 397, 59–63. Jameson, J.C., Erdmann, M.V., 2001. Charting a course toward diagnostic monitoring: a continuing review of coral reef attributes and a research strategy for creating coral reef indexes of biotic integrity. Bull. Mar. Sci. 69 (2), 701–744. Knowlton, N., 2001. Who are the players on coral reefs and does it matter. The importance of coral taxonomy for coral reef management. Bull. Mar. Sci. 69 (2), 305–308. Loya, Y., 1972. Community structure and species diversity of hermatypic corals at Eilat, Red Sea. Mar. Biol. 13 (2), 100– 123.
McKenna, S.A., Richmond, R.H., Roos, G., 2001. Assessing the effects of sewage on coral reefs: developing techniques to detect stress before coral mortality. Bull. Mar. Sci. 69 (2), 517–524. Meesters, E.H., Hilterman, M., Kardinall, E., Keetman, M., de Vries, M., Bak, R.P.M., 2001. Colony size–frequency distributions of scleractinian coral populations: spatial and interspecific variation. Mar. Ecol. Prog. Ser. 209, 43–54. Miller, M.W., Weil, E., Szmant, A.M., 2000. Coral recruitment and juvenile mortality as structuring factors for reef benthic communities in Biscayne National Park. Coral Reefs 19, 115–123. Pennisi, E., 1997. Brighter Prospect for the World’s Coral Reefs? Science 277, 491–493. Pielou, E.R., 1966. Species diversity and pattern-diversity in the study of ecological succession. J. Theor. Biol. 13, 370–383. Risk, M.J., Heikoop, J.M., Edinger, R.V., Erdmann, M.V., 2001. The assessment ‘‘toolbox’’: community-based reef evaluation methods coupled with geochemical techniques to identify sources of stress. Bull. Mar. Sci. 69 (2), 443–458. Sammarco, P.W., 1980. Diadema and its relationship to coral spat mortality: grazing, competition and biological disturbance. J. Exp. Mar. Bio. Ecol. 45, 245–272. Shannon, C.E., Weaver, W., 1948. The Mathematical Theory of Communications. University of Illinois Press, Urbana. p. 117. Shlesinger, Y., Lazar, B., 2001. Artificial swimming lagoon in Eilat: monitoring and management. In: Book of Abstracts of the 33rd Anniversary Conference of the IUI Eilat, pp. 131–132. Simpson, E.H., 1949. Measurements of diversity. Nuatur London 163, 668. Wilkinson, C.R., 2000. Status of Coral Reefs of the World 2000. AIMS, Queensland. Yap, H.T., 1986. Bioindication in coral reef ecosystems. Acta Biol. Hung. 37 (1), 55–58.